With seasonal rainfall predictions having gone wrong twice this year,serious questions are being raised over the Met department and the credibility and accuracy of its long-range forecasts.
The Ministry of Earth Sciences conceded while answering a Parliament question on Wednesday that the accuracy of long-range forecasts in the past four years,excluding this year,has only been about 50 per cent. The short-range forecasts,of the order of a few days,have produced much better results,having an accuracy rate of between 70 and 95 per cent.
The long-range forecasts are made in April of the total amount of rainfall the country is likely to get in the four-month monsoon season from June to September.
What has happened this year has been particularly amusing. On April 26,the Indian Meteorological Department (IMD),in an annual exercise,predicted that the country as a whole would receive rainfall that would be 99 per cent of the long period average (the average of 1941 to 1990),or a normal monsoon season.
But two months later,when it announced the monthly and regional distribution of rainfall,the IMD made a downward revision and said rains were likely to be only 96 per cent of average. June and July,however,saw very less rainfall. By July,the shortfall in rains had reached 23 per cent,at which point the government seemed to panic. In an unusual step,the Prime Ministers Office issued a statement saying the total seasonal rainfall was likely to be only 92 per cent of the average and that the government was making all attempts to deal with a drought-like situation in some parts. A few days later,the IMD made a drastic cut in its estimates and said rainfall was not likely to be more than 85 per cent of average.
But August changed the trend. Persistent rains in most parts of the country ensured that the shortfall came down to 12 per cent. Current indications are that this deficiency might be made up a little further in September and the overall actual rainfall might end up somewhere around 8-10 per cent below average.
All this while,the IMD seems to be reacting to events. By its very nature,weather forecast is an exercise in trying to predict possible outcomes,not react to events after they have happened. The fundamental question therefore is why do we have long-range forecast at all, said a government scientist.
The long-range forecast in April does not tell us when and where will the rainfall occur. There are large variations in rainfall region-wise but that does not get captured in a single average figure for the entire country. For example,none of the IMD predictions till July could see what happened in Rajasthan in August. So we need to very seriously consider whether we should only concentrate on short-range forecasts,which we can do with good accuracy, he said.
D R Sikka,former director of Indian Institute of Tropical Meteorology in Pune,said there were limitations in making long-range forecasts,in particular about extreme events,but these did still have their utility. People will demand it. People want to know how much rain is likely to happen this year. Despite the inaccuracies,there is a keen interest in the seasonal forecast. So we need to keep trying to improve these forecasts, he said.
Dr D Sivananda Pai,who has been making long-range forecasts for the IMD,said incessant rains in Rajasthan in August were a result of intra-seasonal variability in climate conditions,that cannot be captured in any prediction made two months ahead. Such variations can be seen only a few days ahead,he said.